Angle-Integrated LSTM Network in UWB Indoor Localization

To improve the accuracy of UWB localization and solve the weakness of traditional LSTM model in NLOS condition, proposed an Angle-Integrated LSTM Network algorithm to predict the tag positions. The proposed algorithm captures the angle variation during pedestrian movement, enhances the traditional LSTM model by augmenting the parameter dimension of the training. The simulation results show that the proposed algorithm can achieve a nearly 5cm mean localization error as compared to other three algorithms.